Unless you’ve been living under a rock, chances are you’ve heard plenty about ChatGPT over the past year—and with good reason. This generative AI tool, created by OpenAI, lives up to the hype.
In a matter of seconds, it can generate thorough answers to your questions using relevant data. It can summarize large amounts of information, like a call transcription, into an easy-to-read narrative. It can proofread your writing and adjust the style to match the specific tone you are aiming for. And it can even translate languages, including programming languages.
For anyone unfamiliar with the technology, ChatGPT can feel like a magical assistant. But as we know, the reality is that ChatGPT is built on large language models (LLMs) trained on vast amounts of data. This training enables it to make sense of patterns in the data and discover connections in the content it encounters, allowing it to generate original, human-like responses.
These capabilities ultimately led Domo to work with OpenAI and launch our AI Service Layer last year. Our developers were able to easily integrate Domo’s software with OpenAI and quickly build out our own products that help both technical and non-technical Domo users dig deeper into their data and ask better questions. Plus, it complied with our requirements for protecting personal data.
But, we’ve recently switched our LLM to DomoGPT. Currently available in the AI Service Layer, DomoGPT includes a collection of Domo Cloud private models—DomoGPT Chat, Medium; DomoGPT Chat, Small; DomoGPT SQL; and DomoGPT Summarize. These models enable you to continue taking full advantage of the advanced capabilities of AI without sacrificing security.
So what prompted the switch to DomoGPT? In part 5 of our AI Insights Livestream series, we unpacked why this change was necessary.
We are listening to your concerns and securing your data
Since the introduction of the AI Service Layer, the two most frequent questions our team has received from customers are, “What data are you sending to OpenAI?” and “What type of security and data privacy does Domo have?”
To be clear, we have always prioritized our customers’ data security and privacy. But we needed to send metadata and our customer’s queries to OpenAI in order to perform essential tasks. We realized we needed a better solution to ensure that our customers feel confident in Domo’s ability to protect their data and are encouraged to continue using these groundbreaking tools.
We understand the pressure of navigating policy compliance
The explosion of AI tools can be a double-edged sword for professionals. On one hand, you now have more options than ever to harness the capabilities of AI for a range of tasks, including data analysis. On the other hand, companies must thoroughly evaluate these new tools to make sure that they comply with internal policies and requirements.
With developments in AI barreling forward at a rapid pace, companies’ policies can also change quickly as they adapt to new technology. Consequently, professionals may have to give up tools they have already built into their workflows if they are deemed unsafe or unsecure. We are getting ahead of this by making the switch to the more secure DomoGPT so that you can continue your work without disruptions.
We can better maintain data security by keeping your information in the Domo ecosystem
Rather than sending out questions and queries to OpenAI, we knew that the best way to keep your data secure was to ensure that everything stayed in the Domo hosting environment.
With DomoGPT, customer data remains in Domo’s secure private cloud within AWS infrastructure (using Amazon Bedrock). All of Domo’s data protection controls are fully extended to safeguard customer data within this private cloud. And AWS does not log or store customer prompts and completions.
So, when you use AI chat or the AI Service Layer, you can rest assured that your data is not moving unencrypted over the public internet; it stays within the Domo infrastructure.
Your data should be used to train your models and yours alone
But safeguarding your information is just one piece of the puzzle. DomoGPT is ultimately designed to keep your data secure by protecting it from third parties. That doesn’t mean you are losing access to models you’ve built using top tools like Databricks, Amazon, and others that help you perform important tasks with your data.
On the contrary, we recognize how important those models are, which is why we have connectors configured within the AI Service Layer that allow you to quickly and securely connect to and use them within Domo.
What this means is that your data will never be used to train a model unless it’s one that you’ve created and are training yourself. The questions you ask and the data you share will not be accessible to third-party providers for logging or training purposes.
Every interaction with DomoGPT stays entirely within Domo’s private cloud infrastructure, ensuring that your data and queries never leave its secure environment and remain securely isolated from external services or endpoints.
Keeping your most sensitive data secure
At Domo, we believe that AI applications shouldn’t just be easy to use; they also have to be safe. And our teams continue to innovate, ensuring that your experience with the platform remains as secure and transparent as possible.
If you’re ready to make the most of Domo’s AI capabilities, make sure that DomoGPT is available in your instance and that your teams have the appropriate AI Service Layer grants for the individuals who need them.
You can also learn more about the topic by reviewing our frequently asked questions about Domo.AI security and privacy or exploring the resources below.
- Watch the replay of AI Insights: Part 5 – Meet Your Personal Data Assistant, the final part of our summer 2024 AI series.
- Check out our AI Readiness Guide, posted on our Community Forums. It’s a checklist for priming your data sets to be ready for any AI use case.
- Catch up on our best practices for using Domo’s AI Chat agent. If you’re curious about getting more value from AI, this blog is for you.
